Abstract
This paper proposes a framework to analyze hedge funds fee arrangements in which the portfolio construction is determined by the hedge fund manager and the fees are determined via an optimal equilibrium between the manager and the investor. In this setting, fees include management fees (\(\alpha\)) and performance fees (\(\beta\)). We select the paradigm of Expected Utility Theory to determine the managers optimal strategy. Benefiting from the dependence of the optimal terminal payoff on the fee structure, we explore two criteria producing an equilibrium fee mutually satisfying the investor and the manager, one based on Pareto Optimality and the second on negotiable regions. The former also leads to a Pareto efficient frontier of fee structures. We obtain evidence that the popular fee structure of \((\alpha , \beta )=(2\%, 20\%)\) is not an equilibrium fee between manager and investor. Although such equilibrium heavily depends on risk aversion levels and market conditions, the pair \((\alpha , \beta )=(0.5\%, 30.7\%)\) stands out as a fair choice. Moreover, the expected utility of the investor is not monotone in the performance fee for many market conditions. In other words, we prove that fee arrangements which include performance fees are usually beneficial for the investor.
Similar content being viewed by others
Notes
Shifting with \(\alpha X_0\) means that \(\alpha\) has a positive effect on the investor’s wealth and his utility function through the shift. Even so we will see in “Fee preferences” section that the investor wants \(\alpha\) to be minimal.
An alternative approach which allows for \(c_2>1\) would be to use a shifted exponential utility function as \(u_{\mathrm{I}}(X_T)=U_{\mathrm{I}}(I_T(X_T))=1-{\mathrm{e}}^{-c_2(X_T-\beta (X_T-B)^+)}\). However, for the extend of our analysis and for better comparability we focus on the stated HARA utility function.
To obtain greater accuracy without simultaneously blowing up the runtime, we evaluate the value functions \(\phi _{\mathrm{M}}\) and \(\phi _{\mathrm{I}}\) with an increment size of 0.5% for \(\alpha\) and \(\beta\) and use cubic interpolation between those discrete points with an increment size of 0.1%. In our numerical calculations, we furthermore need to discretize \(\varPhi _\mathrm{min}\). We do this with an increment of 0.01.
Please note that the Sharpe ratio is only one possible, but reasonable choice for a potential goal function which we use here for illustrative purposes.
In order to capture the equilibrium fee structure appearing at the kink point, we need to increase the upper bound of in the interval \(\varPhi _\mathrm{min}\) for some scenarios.
References
Berkelaar, Arjan B., Roy Kouwenberg, and Thierry Post. 2004. Optimal portfolio choice under loss aversion. The Review of Economics and Statistics 86: 973–987.
Bichuch, Maxim, and Stephan Sturm. 2014. Portfolio optimization under convex incentive schemes. Finance and Stochastics 18 (4): 873–915.
Buraschi, Andrea, Robert Kosowski, and Worrawat Sritrakul. 2014. Incentives and endogenous risk taking: A structural view on hedge fund alphas. The Journal of Finance 69: 2819–2870.
Carpenter, Jennifer N. 2000. Does option compensation increase managerial risk appetite? The Journal of Finance 55: 2311–2331.
Cox, John C., and Chi-Fu Huang. 1989. Optimal consumption and portfolio policies when asset prices follow a diffusion process. Journal of Economic Theory 491: 33–83.
Cuoco, Domenico, and Ron Kaniel. 2011. Equilibrium prices in the presence of delegated portfolio management. Journal of Financial Economics 101: 2264–296.
Djerroud, Ben, David Saunders, Luis Seco, and Mohammad Shakourifar. 2016. Pricing shared-loss hedge fund fee structures. Innovations in Derivatives Markets 165: 369–383.
Eggertsson, Gauti B. 2010. What fiscal policy is effective at zero interest rates? NBER Macroeconomics Annual 25: 59–112.
Filipovic, Damir, Robert Kremslehner, and Alexander Muermann. 2015. Optimal investment and premium policies under risk shifting and solvency regulation. Journal of Risk and Insurance 82: 261–288.
Goetzmann, William N., Jonathan E. Ingersoll, and Stephen A. Ross. 2003. High-water marks and hedge fund management contracts. The Journal of Finance 58 (4): 1685–1718.
Guasoni, Paolo, and Jan Oblój. 2016. The incentives of hedge fund fees and high-water marks. Mathematical Finance 26 (2): 269–295.
He, Xue D., and Steven Kou. 2016. Profit sharing in hedge funds. Working Paper.
Hodder, James E., and Jens C. Jackwerth. 2007. Incentive contract and hedge fund management. Journal of Financial and Quantitative Analysis 42: 811–826.
Holt, Charles A., and Susan K. Laury. 2002. Risk aversion and incentive effects. American Economic Review 92: 1644–1655.
Karatzas, Ioannis, J. Lehoczky, and Steven E. Shreve. 1987. Explicit solution of a general consumption/investment problem. SIAM Journal of Control and Optimization 25: 1557–1586.
Koijen, Ralph S.J. 2014. The cross-section of managerial ability, incentives, and risk preferences. The Journal of Finance 69 (3): 1051–1098.
Kraft, Holger. 2005. Optimal portfolios and Heston’s stochastic volatility model: An explicit solution for power utility. Quantitative Finance 5: 303–313.
Lan, Yingcong, Neng Wang, and Jinqiang Yang. 2013. The economics of hedge funds. Journal of Financial Economics 110 (2): 300–323.
Merton, Robert C. 1969. Lifetime portfolio selection under uncertainty: The continuous-time case. Review of Economics and Statistics 51: 247–257.
Merton, Robert C. 1971. Optimum consumption and portfolio rules in a continuous-time model. Journal of Economic Theory 51: 373–413.
Morgenstern, Oskar, and John Von Neumann. 1953. Theory of games and economic behavior. Princeton university press.
Puri, Karan, Andreas Krause. 2015. Optimal hedge fund contract parameters: Theory and empirical evidence. Working paper. https://editorialexpress.com/cgi-bin/conference/download.cgi?db_name=MMF2015&paper_id=176. Accessed 12 Nov 2018.
Reichlin, Christian. 2013. Utility maximization with a given pricing measure when the utility is not necessarily concave. Mathematics and Financial Economics 7 (4): 531–556.
Stabile, Gabriele. 2015. Underperformance fees and manager’s portfolio risk taking. International Journal of Financial Research 6: 79–89.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Figures
See Figs. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 and 25.
Appendix 2: Mathematical proofs
Proof of Lemma1:
We show that (13) is equivalent to (15). Therefore, we use the following result.
with
Hence,
By application to (13), this leads to
as presented in Lemma 1. \(\square\)
Proof of Proposition 2:
The proposition follows with \({\mathbb {V}}\left[ X_T^*\right] ={\mathbb {E}}\left[ \left( X_T^*\right) ^2\right] - \left( {\mathbb {E}}\left[ X_T^*\right] \right) ^2\). \(\square\)
Proof of Lemma 2:
Notice that \(x^*>B\) and \(\lambda ^*{\tilde{Z}}_T<v_{\mathrm{M}}'(x^*)\iff X_T^*>x^*.\) Hence, we can calculate
\(\square\)
Proof of Lemma 3:
Let \(\alpha , \beta\) and the optimal payoff \(X_T(\alpha ,\beta )\) be given. Now, let us choose \((\alpha ',\beta ')>(\alpha ,\beta )\), but the same \(X_T(\alpha ,\beta )\). Then, the budget constraint is fulfilled and
After taking expectations and then optimizing the left-hand side in \(X_T\) to find the optimal \(X_T(\alpha ',\beta ')\), we learn that the left-hand side must increase, giving us \(\phi _{\mathrm{M}}(\alpha ',\beta ')>\phi _{\mathrm{M}}(\alpha ,\beta )\). \(\square\)
Proof of Proposition 3
Let \(\alpha >0\) be arbitrary, but fixed. The main idea of the proof is to show that for a small \(\beta _2\) and \(\beta _1\rightarrow 0\) the optimal payoff \(X_{\beta _1}^*\) is first-order stochastically dominated by \(X_{\beta _2}^*\) which means that \({\mathbb {E}}[u(X_{\beta _1})]\le {\mathbb {E}}[u(X_{\beta _2})]\) for all increasing utility functions u. The proof is structured in six parts. We show:
-
(1)
\(x_{\beta }^*\) is a decreasing and \(v_{\mathrm{M}}'(x_{\beta }^*)\) is an increasing function in \({\beta }\).
-
(2)
\(\lambda _{\beta }^*\) is an increasing function in \(\beta\).
-
(3)
\(\frac{v_{\mathrm{M}}'(x_{\beta }^*)}{\beta }\) converges to a constant for \(\beta \rightarrow 0\).
-
(4)
\(\frac{\lambda _{\beta }^*}{\beta }\) and \(\frac{\lambda _{\beta }^*}{v_{\mathrm{M}}'(x_{\beta }^*)}\) are decreasing functions in \(\beta\), for small \(\beta\).
-
(5)
Estimates for help functions.
-
(6)
Stochastic dominance of \(X_{\beta _2}^*\) over \(X_{\beta _1}^*\), for small \(\beta _2\) and \(\beta _1\rightarrow 0\).
(1) \(x_{\beta }^*\) is a decreasing and \(\mathbf {v_{{M}}'}{(\mathbf {x}_{\beta }^*)}\) is an increasing function in \({\beta }\).
Let us define \(u_{\mathrm{M}}^{\beta }(x)=\frac{1}{1-c_1}\left( \alpha X_0+\beta (x-B)^+\right) ^{1-c_1}\), \(x>0\) arbitrary. It follows:
-
(i)
\(u_{\mathrm{M}}^{\beta _1}(x)<u_{\mathrm{M}}^{\beta _2}(x) \quad \forall \beta _1<\beta _2, \forall x>B\)
-
(ii)
\(\frac{\partial }{\partial x}u_{\mathrm{M}}^{\beta _1}(x)>\frac{\partial }{\partial x}u_{\mathrm{M}}^{\beta _2}(x) \quad \forall \beta _1<\beta _2, \forall x>B\)
-
(iii)
\(\frac{\partial }{\partial x}u_{\mathrm{M}}^{\beta }> 0\quad \forall x>B\)
-
(iv)
\(\frac{\partial ^2}{\partial ^2 x}u_{\mathrm{M}}^{\beta }(x)< 0\quad \forall x>B\)
Let \(\beta _1<\beta _2\). From the definition of \(x^*\), we know:
First, we show that \(x_{\beta }^*\) is decreasing in \(\beta\). We assume \(x_{\beta _1}^*<x_{\beta _2}^*\). Then, it is
which is a contradiction. Hence, it must be \(x_{\beta _1}^*>x_{\beta _2}^*\), i.e., \(x_{\beta }^*\) is decreasing in \(\beta\). Let \(v_{\mathrm{M}}\) be the concavification of \(u_{\mathrm{M}}^{\beta }\). Since the concavification \(v_{\mathrm{M}}\) is a concave function, it follows \(v_{\mathrm{M}}'(x_{\beta _2}^*)>v_{\mathrm{M}}'(x_{\beta _1}^*)\), such that \(v_{\mathrm{M}}'(x_{\beta }^*)\) is an increasing function of \(\beta\).
(2) \(\lambda _{\beta }^*\) is an increasing function in \(\beta\).
We look at Eq. (15):
The right-hand side of the upper equation is constant while the left-hand side appears to be increasing in \(\beta\) (remember that \(v_{\mathrm{M}}'(x_{\beta }^*)\) is increasing in \(\beta\)) for \(c_1<1\) (while assuming \(\lambda _{\beta }^*\) to be constant) and simultaneously decreasing in \(\lambda _{\beta }^*\) (while assuming \(\beta\) to be constant). Therefore, \(\lambda _{\beta }^*\) is increasing in \(\beta\). This implies that \(\frac{\lambda _{\beta }^*}{\beta }\) is monotone for small \(\beta\) which is relevant for Step 4).
(3) \({\frac{{\mathbf {v_M'}}(x_{\beta }^*)}{\beta }}\) converges to a constant for \(\beta {\rightarrow }\) 0.
Here we show that \(v_{\mathrm{M}}'(x_{\beta }^*)\sim \beta\), which means \(\lim \nolimits _{\beta \rightarrow 0}\frac{v_{\mathrm{M}}'(x_{\beta }^*)}{\beta }=a\) where \(a>0\) is a constant.
From (12), the definition of \(x_{\beta }^*\), we know
For \(\beta \rightarrow 0\), there are three different possible cases for \(\beta x_{\beta }^*\).
-
\(\beta x_{\beta }*\rightarrow \infty\)
-
\(\beta x_{\beta }*\rightarrow 0\)
-
\(\beta x_{\beta }*\rightarrow {\hbox {const.}}>0\)
For the first two cases, we differentiate the left side of the upper equation w.r.t. \((\beta x_{\beta }*)\) and apply the rule of l’Hopital. Hence, for \(\beta \rightarrow 0\) it must hold
-
For \(\beta x_{\beta }*\rightarrow \infty\), this leads to \(1\ne 1-c_1\).
-
For \(\beta x_{\beta }*\rightarrow 0,\) this leads to \(1-c_1=1-c_1\). \(\checkmark\)
The third case, \(\beta x_{\beta }*\rightarrow {\hbox {const.}}>0\), cannot be true, because the unique limit cannot be 0 and a constant greater than 0 simultaneously. Therefore, \(\lim \nolimits _{\beta \rightarrow 0}\beta x_{\beta }^*= 0\).
Since \(v_{\mathrm{M}}'(x_{\beta }^*)=\beta \cdot (\alpha X_0+\beta (x_{\beta }^*-B))^{-c_1}\), we obtain
where \((\alpha X_0)^{-c_1}\) is constant. Hence, \(v_{\mathrm{M}}'(x_{\beta }^*)\sim \beta\) as prescribed.
(4) \({\frac{{\varvec{\lambda }}_{\beta }^*}{\beta }}\) and \({\frac{{\varvec{\lambda }}_{\beta }^*}{{\mathbf {v_M}}'({\mathbf {x}}_{\beta }^*)}}\) are decreasing functions in \({\beta }\), for small \({\beta }\).
We define \(y_{\beta }=\frac{\lambda _{\beta }^*}{\beta }\) and use the result from Step 3)
Furthermore, let
and
Then, using (15) we get
Looking at Eq. (37) and recalling that \(y_{\beta }\) is monotone for small \(\beta\), we obtain:
-
Assume \(y_{\beta }\downarrow 0\): Then, \(\lim \nolimits _{\beta \rightarrow 0}L(\beta )=E_3\cdot \infty -\alpha X_0\cdot E_4\ne 0=\lim \nolimits _{\beta \rightarrow 0}R(\beta )\Rightarrow\) Contradiction to (37)
-
Assume \(y_{\beta }\uparrow \infty\): Then, \(\lim \nolimits _{\beta \rightarrow 0}L(\beta )=E_3\cdot 0-0=0=\lim \nolimits _{\beta \rightarrow 0}R(\beta )\) \(\checkmark\)
Analogously to the argumentation in Step 3, \(y_{\beta }\rightarrow y^*>0\) cannot hold, because \(y_{\beta }\) cannot converge to \(\infty\) and a constant greater than 0 simultaneously. Hence, \(\lim \nolimits _{\beta \rightarrow 0}y_{\beta }=\infty\), and thus, for small \(\beta\), \(\frac{\lambda _{\beta }^*}{\beta }\) and \(\frac{\lambda _{\beta }^*}{v_{\mathrm{M}}'(x_{\beta }^*)}\) are decreasing in \(\beta\).
(5) Estimates for help functions.
Let \(f(\beta ):=\left( \frac{\lambda _{\beta }^*}{\beta }\right) ^{-1/c_1}\cdot \frac{1}{\beta }\). Using Eq. (15) and the fact that \(\frac{\lambda _{\beta }^*}{v_{\mathrm{M}}'(x_{\beta }^*)}\) is decreasing in \(\beta\) for small \(\beta\), the terms \(\left[ 1-N\left( d\left( \frac{\lambda _{\beta }^*}{v_{\mathrm{M}}'(x_{\beta }^*)}\right) +\varepsilon _1\right) \right]\), \(\left[ 1-N\left( d\left( \frac{\lambda _{\beta }^*}{v_{\mathrm{M}}'(x_{\beta }^*)}\right) +\varepsilon _2\right) \right]\) and \(\left( B-\frac{\alpha X_0}{\beta }\right)\) are increasing in \(\beta\). It follows that \(f(\beta )\) is decreasing in \(\beta\) for small \(\beta\).
Let \(\beta _1<\beta _2\) be small and define \(g(\beta _1,\beta _2):=\left( \frac{\lambda _{\beta _2}^*}{\lambda _{\beta _1}^*}\cdot \frac{\beta _1}{\beta _2}\right) ^{1/c_1}\). Then, it holds
-
(I)
\(g(\beta _1,\beta _2)<1\), as direct result of \(\frac{\lambda _{\beta _2}^*}{\beta _2}<\frac{\lambda _{\beta _1}^*}{\beta _1}\).
-
(II)
\(\beta _1-\beta _2\cdot g(\beta _1,\beta _2)<0\iff \frac{\beta _1}{\beta _2}<g(\beta _1,\beta _2)\) which follows from \(f(\beta _2)<f(\beta _1)\).
(6) Stochastic dominance of \(X_{\beta _2}^*\) over \(X_{\beta _1}^*\), for small \(\beta _2\) and \(\beta _1\rightarrow 0\).
In the final step, we now look at the cumulative distribution function of the optimal payoff
Remember that \(X_{\beta }^*\) never takes values between 0 and \(x_{\beta }^*>B\). Let \(x>x_{\beta }^*\) be arbitrary. Then:
Hence, for any \(x>x_{\beta }^*\) the cumulative distribution function is given by
Let \(\beta _1<\beta _2\) be small and \(x\ge x_{\beta _1}^*>x_{\beta _2}^*\) (where the last inequality comes from Step 1). Then,
Now we show that \(S(\beta _1,\beta _2)\rightarrow \infty\) for \(\beta _1\rightarrow 0\).
In Step 4, we have seen that \(\frac{\lambda _{\beta }^*}{\beta }\rightarrow \infty\) for \(\beta \rightarrow 0\). Hence, \(\frac{\beta _1}{\lambda _{\beta _1}^*}\downarrow 0\) for \(\beta _1\rightarrow 0\). Therefore, we know
(I) and (II) in Step 5 imply that \(-\beta _2+\frac{\beta _2-\beta _1}{1-g(\beta _1,\beta _2)}\) is always positive. Thus, Eq. (38) converges to 0 from above for \(\beta _1\rightarrow 0\), such that \(S(\beta _1,\beta _2) \rightarrow \infty\) for \(\beta _1\rightarrow 0\) and small \(\beta _2\).
Hence, for \(\beta _1\rightarrow 0\) and small \(\beta _2\) it holds
i.e., \(X_{\beta _2}^*\) dominates \(X_{\beta _1}^*\) on \(\varOmega _1\) by first-order stochastic dominance. Therefore,
We thus conclude
\(\square\)
Proof of Lemma 4:
Let \(({\hat{\alpha }},{\hat{\beta }})\notin P\). Set \(\phi _\mathrm{min}= \phi _{\mathrm{M}}({\hat{\alpha }},{\hat{\beta }})\) and let \((\alpha ,\beta )\) be the solution of the corresponding problem (28). As \(({\hat{\alpha }},{\hat{\beta }})\) is admissible for (28) and \((\alpha ,\beta )\) the optimal solution, we conclude
As \(({\hat{\alpha }},{\hat{\beta }})\notin P\), we know that \(({\hat{\alpha }},{\hat{\beta }})\) cannot be an optimal solution for (28). Hence, the inequality must be strict. \(\square\)
Rights and permissions
About this article
Cite this article
Escobar-Anel, M., Höhn, V., Seco, L. et al. Optimal fee structures in hedge funds. J Asset Manag 19, 522–542 (2018). https://doi.org/10.1057/s41260-018-0094-7
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1057/s41260-018-0094-7